High Sentiment Research Articles

Short Selling Threat and News Sentiment Management in Press Releases

This study presents evidence that firms actively manage sentiment in press releases to reduce their exposure to short selling threats. I exploit the suspension of short selling constraints as an exogenous shock to the short selling threat borne by firms. Using a sample of 110,938 firm-press release observations from 2001 to 2007 and a difference-in-differences methodology, I find that firms facing increased short selling threats increase the sentiment in press releases after the suspension of short selling constraints more than otherwise similar firms without increased short selling threats. Meanwhile, higher sentiment in corporate press releases deters short sellers. Specifically, stocks with the suspension of short selling constraints are more likely to be short sold but such stocks with higher news sentiments are less likely to be short sold. These findings suggest that press release sentiment can be strategically used by firms to advance their own interests.

Investor sentiment and stock market anomalies: Evidence from Islamic countries

Studies of the Ramadan effect argue that higher stock returns in Muslim countries during Ramadan relate to higher investor sentiment. However, Islamic countries rank low on the Hofstede Individualism Index, a proxy for investor overconfidence. Therefore, this study examines the impact of investor sentiment on stock market anomalies in two advanced Islamic finance jurisdictions: Malaysia and Indonesia. It hypothesizes that stock market anomalies are stronger following high sentiment if investors in Malaysia and Indonesia are overconfident. The results show that the long and short legs of the stock market anomalies earn relatively low returns following high investor sentiment, indicating overpricing during high sentiment. Moreover, the short leg earns relatively lower returns than the long leg following high sentiment because the short leg is more overpriced than the long leg when sentiment is high. Therefore, consistent with the hypothesis, the long-short returns of anomalies are stronger following high investor sentiment because of the relatively lower returns of the short leg than the long leg.

Certainty of uncertainty for asset pricing

Uncertainty is known to be crucial in asset pricing, yet evidence from a comprehensive analysis of various uncertainty measures remains sparse. By machine learning, we construct a novel economic uncertainty index derived from a heterogeneous range of uncertainty measures and investigate its predictability of stock returns. Our composite uncertainty index exhibits robust in- and out-of-sample predictability of stock market returns over the one- to 12-month horizon. The predictive power stems from the volatility-orthogonal components of individual uncertainty measures and becomes more pronounced during high uncertainty and high sentiment periods. The predictability of our economic uncertainty index aligns with theoretical frameworks linking uncertainty to future investment, cash flows, and market expectations.

SentiGEN: Synthetic Data Generator for Sentiment Analysis

Obtaining high-quality, diverse, accurate datasets for sentiment analysis has always been a significant challenge. Traditional approaches include annotators, which may introduce bias to datasets and are also time-consuming and expensive. These types of datasets may also not represent the variety needed to train robust and generalizable sentiment analysis models. This study introduces a novel combination of techniques to approach the problem with a novel solution. The proposed system, SentiGEN includes the use of a transformer, T5, fine-tuned and optimized using an evolutionary algorithm to generate high-quality, diverse, accurate data for sentiment analysis. The generated data is validated using XLNet to ensure high sentiment accuracy. This combination of technologies has proven successful based on the results derived from evaluating multiple models. From complex transformers such as BERT to more straightforward approaches like KNN, those trained using synthetic data demonstrated superior performance compared to their counterparts trained on real data. This enhancement in predictive accuracy was observed when evaluated on benchmark datasets such as SST-2 and Yelp. SentiGEN can generate high-quality, diverse, accurate, realistic data for sentiment analysis and successfully increased the performance of models trained on synthetic data compared to the same model trained on real data.

Timing sentiment with style: Evidence from mutual funds

This study analyzes four distinct types of mutual fund sentiment timing skills using a multifactor framework. Our results indicate a diminished significance of market sentiment timing, in contrast to the results of prior studies. Additionally, we reveal that size and value sentiment timing can substantially enhance fund performance. Managers strategically reduce their exposure to small stocks using size sentiment timing and increase exposure to value stocks through value sentiment timing during high sentiment periods. We find no evidence that mutual fund managers engage in momentum sentiment timing.

State-dependent intertemporal risk-return tradeoff: Further evidence

We suggest that the intertemporal risk-return tradeoff is not necessarily positive but rather state dependent. We further explore the state dependent risk-return relation by examining how the positive risk-return relation is distorted in response to various market conditions, including extreme price changes, differing levels of investor sentiment, the introduction of stock options, and throughout business cycles. The tendency for uninformed investors to be optimistic (pessimistic) in response to good (bad) market news cause overpricing (underpricing), and the resulting trade activity of arbitrageurs that distorts the positive risk-return tradeoff, is documented consistently across these environments. We find that the attenuation (reinforcement) of the positive risk-return relation under investors’ optimistic (pessimistic) expectations is stronger in high (low) sentiment periods, in the presence of extreme returns, in the period after stock options became available, and during expansionary periods. We argue that the asymmetric intertemporal risk-return relation is a consequence of rational arbitrageurs’ trading to exploit mispricing through the selling of overpriced stocks conditional on good news and buying underpriced stocks conditional on bad news.

Institutional herding and investor sentiment

We investigate the role of investor sentiment in institutional herding behavior and its impact on stock prices. We find that institutional investors exhibit more herding behavior during periods of high sentiment, which has a significant impact on stock prices. Our results show that herding has a stabilizing effect on the stock market when investor sentiment is low, while it causes price distortions when sentiment is high. We also show that the impact of sentiment on price is particularly pronounced for small, non-profitable, low tangibility, high-growth firms.

Manager sentiment and conditional conservatism

AbstractThis study examines the effect of manager sentiment on conditional conservatism. Manager sentiment refers to widely held beliefs of financial managers about their firms’ future economic prospects that are not justified by available economic fundamentals. Manager sentiment is likely to affect conditional conservative reporting because the decision to recognize unrealized economic losses in a timely manner flows from financial managers’ beliefs about their firms’ future cash flow prospects. We predict and find that manager sentiment is negatively associated with conditional conservatism, indicating that firms report less conservatively during periods of high manager sentiment (over‐optimism) and more conservatively during periods of low manager sentiment (over‐pessimism). Moreover, the effects of manager sentiment on conditional conservatism remain strongly negative after controlling for manager overconfidence. We further find that asset write‐offs are lower during high sentiment periods but higher in subsequent periods. Importantly, the manager sentiment effect on conservatism is incremental, and the opposite in sign, to the effect of investor sentiment, which has not been demonstrated in prior literature.

Geographic links and predictable returns

AbstractUsing establishment‐level data of U.S. public firms, we construct a novel measure of geographic linkage between firms. We show that the returns of geography‐linked firms have strong predictive power for focal firm returns and fundamentals. This effect is distinct from other cross‐firm return predictability and is not easily attributable to risk‐based explanations. It is more pronounced for focal firms that receive lower investor attention, are more costly to arbitrage, and during high sentiment periods. The cross‐firm information spillovers and return predictability are also stronger for geographic peers with economic linkages and with positive information. Our results are broadly consistent with sluggish price adjustment to nuanced geographic information.

The impact of narration, sentiment, and emotional modulation on economic of decision-making: a multidisciplinary study

The article delves into the interplay between narration, sentiment, and decision-making processes, particularly within the realm of economic choices. Emphasizing the pivotal role of message content, the authors present a hypothesis depicting the profound impact of narration, characters, and sentiment in the communication process. These components wield significant influence over the emotional facets of decision-making, influencing the process or prompting alterations to initial decisions. Critically challenging established economic paradigms, the article underscores the indispensable emotional dimension inherent in communication. Drawing upon the scholarly contributions of eminent figures like Robert J. Shiller, the authors synthesize insights from psychology and neurobiology, unraveling the manipulative essence inherent in narration. In the theoretical section, the article also references Uri Hasson's theories and research on brain synchronization during the communication process. The practical segment of the article substantiates these claims through the presentation of findings from neuropsychological studies. The study confirms the hypothesis that narratives infused with high sentiment resulting from emotional interactions, following a cause-and-effect structure with a visible beginning, development, and end, lead to better synchronization and intensify neuronal responses. The structure, plot shaping, and sentiment in messages, influenced by word choice, shape narrative economy and impact cognitive costs in decision-making. The obtained results confirm that precisely tailored narration effectively reduces cognitive costs associated with the decision-making process. This validates the authors' hypothesis, opening up space for research replication and falsification of the results obtained.

Identifying self-disclosed anxiety on Twitter: A natural language processing approach

BackgroundText analyses of social media posts are a promising source of mental health information. This study used natural language processing to explore distinct language patterns on Twitter related to self-reported anxiety diagnosis. MethodsA total of 233.000 tweets made by 605 users (300 reporting anxiety diagnosis and 305 not) over six months were comparatively analysed, considering user behavior, Linguistic Inquiry Word Count (LIWC), and sentiment analysis. Twitter users with a self-disclosed diagnosis of anxiety were classified as ‘anxious’ to facilitate group comparisons. ResultsSupervised machine learning models showed a high prediction accuracy (Naïve Bayes 81.1 %, Random Forests 79.8 %, and LASSO-regression 79.4 %) in identifying Twitter users’ self-disclosed diagnosis of anxiety. Additionally, a Latent Profile Analysis (LPA) identified four profiles characterized by high sentiment (31 % anxious participants), low sentiment (68 % anxious), self-immersed (80 % anxious), and normative behavior (38 % anxious). ConclusionThe digital footprint of self-disclosed anxiety on Twitter posts presented a high frequency of words conveying either negative sentiment, a low frequency of positive sentiment, a reduced frequency of posting, and lengthier texts. These distinct patterns enabled highly accurate prediction of anxiety diagnosis. On this basis, appropriately resourced, awareness raising, online mental health campaigns are advocated.

Are markets sentiment driving the price bubbles in the virtual?

This paper investigates the existence of speculative bubbles in four crypto-market components (Bitcoin, Ethereum, CRIX index, DeFi pulse index) while date-stamping them, before using a wavelet coherence approach to look after the co-movement of each of those four digital tokens with three sentiment indices reflecting three different markets (Stock, gold, and crypto markets). Thereafter, we study the sentiment-price return connection and its concordance with periods of existing bubbles. Our results show the existence of multiple bubble regimes in the virtual market and more specifically, a huge bubble occurring during the crypto-market bull phase for Bitcoin, Ethereum, and CRIX. Moreover, we find that the bubbles that occurred during the COVID-19 pandemic concord with periods of high sentiment co-movement with the three markets. Hence, some of those bubble periods are in phases with periods of high co-movement with sentiments toward different markets. In addition, the fear and greed toward the crypto-market seem to be the leading sentiment index that highly influences the crypto-market components and their bubble appearances.

Effect of sectoral holdings on the flow-performance sensitivity of mutual funds

We find that the flow-performance sensitivity (FPS) of mutual funds depends on the composition of their sectoral holdings. We use the Morningstar classification of fund holdings into the following three Super Sectors: Defensive, Sensitive, and Cyclical. On average, the FPS decreases as the fraction of defensive or sensitive stocks increases in the fund’s portfolio. The FPS increases as the fraction of cyclical stocks increases. During high sentiment periods, the sensitivity of new sales as well as redemptions increases, resulting in overall higher FPS for all funds. However, in both the low and the high sentiment periods, the FPS is lower for funds with a higher fraction of defensive or sensitive stocks and higher for funds with a higher fraction of cyclical stocks. Investors with a long investment horizon may wish to avoid mutual funds that invest primarily in cyclical stocks as funds with high FPS tend to have high liquidity costs.

The Impact of Technology Attention and Investors on Stock Returns: Taking Science and Technology Innovation Board Company as an Example

Since the fourth anniversary of the opening of the Science and Technology Innovation Board, the number of listed companies on the board has reached 546, with a total fundraising amount of 858.242-billion-yuan, accounting for 42% of the total domestic A-share financing during the same period. The stock return on the Science and Technology Innovation Board has become a topic of concern for investors. This article uses indirect indicators from 2021 to 2023, and uses principal component analysis to construct investor attention and investor sentiment indicators. By using panel analysis, it is found that investor attention has a negative impact on stock return, investor sentiment has a positive moderating effect on investor attention-stock return. In order to further explore the impact and effect of differences in investor sentiment and investor attention in the high and low sentimental group, it conducted a heterogeneity test, which found that the relationship between investor attention and stock returns in the low sentiment group is significantly negative, but not in high group. And there is a significant positive correlation between the investor sentiment moderating effect and stock returns in the high sentiment group, but not in the low sentiment group. This result can be beneficial for guiding market regulatory authorities and consumers to supervise and invest rationally in stocks on the Science and Technology Innovation Board.

Machine-learning the skill of mutual fund managers

We show, using machine learning, that fund characteristics can consistently differentiate high from low-performing mutual funds, before and after fees. The outperformance persists for more than three years. Fund momentum and fund flow are the most important predictors of future risk-adjusted fund performance, while characteristics of the stocks that funds hold are not predictive. Returns of predictive long-short portfolios are higher following a period of high sentiment. Our estimation with neural networks enables us to uncover novel and substantial interaction effects between sentiment and both fund flow and fund momentum.

Sentiment analysis of epidemiological surveillance reports on COVID-19 in Greece using machine learning models.

The present research deals with sentiment analysis performed with Microsoft Azure Machine Learning Studio to classify Facebook posts on the Greek National Public Health Organization (EODY) from November 2021 to January 2022 during the pandemic. Positive, negative and neutral sentiments were included after processing 300 reviews. This approach involved analyzing the words appearing in the comments and exploring the sentiments related to daily surveillance reports of COVID-19 published on the EODY Facebook page. Moreover, machine learning algorithms were implemented to predict the classification of sentiments. This research assesses the efficiency of a few popular machine learning models, which is one of the initial efforts in Greece in this domain. People have negative sentiments toward COVID surveillance reports. Words with the highest frequency of occurrence include government, vaccinated people, unvaccinated, telephone communication, health measures, virus, COVID-19 rapid/molecular tests, and of course, COVID-19. The experimental results disclose additionally that two classifiers, namely two class Neural Network and two class Bayes Point Machine, achieved high sentiment analysis accuracy and F1 score, particularly 87% and over 35%. A significant limitation of this study may be the need for more comparison with other research attempts that identified the sentiments of the EODY surveillance reports of COVID in Greece. Machine learning models can provide critical information combating public health hazards and enrich communication strategies and proactive actions in public health issues and opinion management during the COVID-19 pandemic.

China's diversification discount: The role of the information environment

Based on information uncertainty, in this paper, we provide a mispricing explanation for the diversification discount in the Chinese A-shares market. The empirical evidence reveals an overestimation of a diversified company's value due to the uncertain information of such companies, and this overestimation is stronger when exacerbated by high sentiment. Moreover, we provide new evidence of the year breakpoint in the diversification discount during different sample subintervals, which relies on a change in ex post mispricing related to the macroeconomic environment.

Investor sentiment and firm capital structure

We provide novel evidence of the role of investor sentiment in determining firms' capital structure decisions from three perspectives: leverage ratio, debt maturity and leverage target adjustment. We find that when investor sentiment is high, firms increase their leverage ratios, supporting our contention that high investor sentiment increases firms' debt capacity and facilitates the use of an aggressive leverage policy. Debt maturity is shorter in high sentiment periods, implying that firms are confident about future earnings and use shorter debt maturity to signal their financial solvency. Leverage target adjustment is slower in low sentiment periods, indicating higher costs of external finance. Furthermore, the sentiment-leverage relationship sensitivity is greater for financially constrained firms. Our extended analysis determines that leverage-increasing firms generate lower stock returns subsequent to a period of high sentiment, offering practical insights into the economic consequences of increasing leverage in high sentiment periods on corporate value for investors. Our research advances the understanding of the impact of investor sentiment on firms' financing decisions and stock returns.

Sentiment Analysis of Celebrity-endorsed Advertisements

The usage of Celebrity endorsers in brand advertising has become popular since the presence of a celebrity endorser is likely to positively influence the consumer to buy their products. Marketers expect that the celebrity’s endorsement will create an emotional tie to the product/brand and his/her good attributes will be transferred to the brand. Our research has focused on the impact of celebrity advertisements on consumers using sentiment analysis. Our objective was to use sentiment analysis to calculate the sentiment scores of advertisements with celebrity endorsements. We used four celebrity-endorsed advertisements and collected responses/reactions from consumers. The advertisement responses were analysed using ‘sentimentr’ for endorser congruence and endorser frequency conditions. The findings revealed that brand-endorser congruence and highly frequently appeared endorsers have received high sentiment scores. This research has shown that marketers need to choose brand-congruent and highly frequent endorsers for creating advertising impact.

Investor Global Sentiment and Stock Returns Connectedness in Developed Markets

This article investigates the dynamic connectedness between equity returns and investor sentiment indexes in developed markets. The author uses a network methodology to measure stock return connectedness and investor sentiment spillover. The study examines the role of global investor sentiment in affecting stock return connectedness by employing a quantile-on-quantile approach using data from six developed markets ranging from January 1991, to December 2020. The results show that both returns and sentiment are interconnected and that spillovers across the stock indexes were more pronounced during the Global Financial Crisis of 2008. Moreover, the results suggest that global investor sentiment contributes negatively to the connectivity of stock market returns for the central quantiles of sentiment. However, the relationship becomes positive for extremely high sentiment values.